Our goals are to determine the cognitive and neural processes that comprise different kinds of cognitive control in normal, and to specify how these processes are compromised in patients with schizophrenia, leading to important behavioral deficits. Prior studies of cognitive-control deficits in schizophrenia are limited in wht they can tell us about the specific neural circuits involved because the studies typically use a single task that requires multiple control processes. To some extent, the same limitation is true of prior studies with normal. We propose a different strategy: while having their brains scanned using functional Magnetic Resonance Imaging (fMRI), we will have patients with schizophrenia and normal controls perform multiple cognitive control tasks, where each task is relatively process-pure but taken together the tasks span much of the domain of cognitive control. We will use three state-of-the-art cognitive-neuroscience tasks that based on previous findings with normal should isolate specific control processes and their underlying neural circuitry. The tasks are all variants of the Item-Recognition test of working memory, and include the: (1) Ignore task, in which the participant selectively ignores information before it enters working memory; (2) Suppress task, in which the participant selectively suppresses information that is already in working memory; and (3) Stop task, in which the participant inhibits an already prepared response to a probe of working memory. Over the course of three fMRI sessions, we will test 24 patients with schizophrenia and 24 controls on all three tasks. The behavioral data will provide a precise description of the degree to which each of three different control processes--perceptual selection, working-memory selection, and response inhibition--are intact or impaired in schizophrenia. The fMRI data should reveal three overlapping neural networks that underlie normal cognitive- control processes, as well as inform us about the circuits that are compromised in those control tasks in which the patients show behavioral deficits. Our preliminary behavioral findings indicate that the patients are impaired on Suppress and Stop but not on Ignore--indicating that they do not have a general deficit in cognitive control. Our preliminary fMRI data indicate that the patients' impairment on Suppress is accompanied by abnormal activation (not merely less activation) in the left-hemisphere prefrontal cortex. We expect that our fMRI experiments will further reveal that: while patients show abnormal activity in Suppress, their neural activity in Ignore may be relatively normal; and while normal activate a left PFC circuit in Suppress, they activate a right- hemisphere circuit in Stop, and both these circuits show abnormal activity in patients with schizophrenia.